Vacouver Street Trees

Exploratory Data Analysis | Pankti Shah | December 2021

---

---

Exploratory Data Analysis

# Importing in your required libraries
import pandas as pd
import altair as alt
alt.data_transformers.enable('default', max_rows=1000000)
import json

# Read in the data
url = 'https://raw.githubusercontent.com/UBC-MDS/data_viz_wrangled/main/data/Trees_data_sets/small_vancouver_trees.csv'
df = pd.read_csv(url, parse_dates=['date_planted'])
df.head()

	Unnamed: 0	std_street	on_street	species_name	neighbourhood_name	date_planted	diameter	street_side_name	genus_name	assigned	...	plant_area	curb	tree_id	common_name	height_range_id	on_street_block	cultivar_name	root_barrier	latitude	longitude
0	19886	W 10TH AV	W 10TH AV	BIGNONIOIDES	Kitsilano	NaT	34.0	ODD	CATALPA	N	...	10	Y	9945	COMMON CATALPA	5	3200	NaN	N	49.263400	-123.177100
1	7941	W 59TH AV	W 59TH AV	SACCHARINUM	Marpole	NaT	20.0	ODD	ACER	Y	...	16	Y	50427	SILVER MAPLE	4	700	NaN	N	49.217059	-123.120787
2	4613	W 47TH AV	W 47TH AV	PLATANOIDES	Kerrisdale	NaT	24.0	ODD	ACER	N	...	12	Y	43456	NORWAY MAPLE	5	2200	NaN	N	49.229119	-123.159841
3	7388	COMMERCIAL DRIVE	COMMERCIAL DRIVE	EUCHLORA X	Grandview-Woodland	NaT	8.0	EVEN	TILIA	N	...	C	Y	69099	CRIMEAN LINDEN	3	1300	NaN	N	49.272647	-123.069463
4	1894	E 55TH AV	E 55TH AV	SPECIES	Victoria-Fraserview	NaT	14.0	EVEN	ABIES	N	...	B	Y	164752	CRIMSON SUNSET NORWAY MAPLE	5	1900	NaN	N	49.219958	-123.067159

5 rows × 21 columns

df.describe(include='all')

<ipython-input-13-174ba9bf1a5c>:1: FutureWarning: Treating datetime data as categorical rather than numeric in `.describe` is deprecated and will be removed in a future version of pandas. Specify `datetime_is_numeric=True` to silence this warning and adopt the future behavior now.
  df.describe(include='all')

	Unnamed: 0	std_street	on_street	species_name	neighbourhood_name	date_planted	diameter	street_side_name	genus_name	assigned	...	plant_area	curb	tree_id	common_name	height_range_id	on_street_block	cultivar_name	root_barrier	latitude	longitude
count	5000.000000	5000	5000	5000	5000	2338	5000.000000	5000	5000	5000	...	4963	5000	5000.000000	5000	5000.000000	5000.000000	2700	5000	5000.000000	5000.000000
unique	NaN	598	607	157	22	1466	NaN	3	62	2	...	41	2	NaN	339	NaN	NaN	164	2	NaN	NaN
top	NaN	W KING EDWARD AV	W KING EDWARD AV	SERRULATA	Kensington-Cedar Cottage	2006-11-21 00:00:00	NaN	ODD	ACER	N	...	10	Y	NaN	KWANZAN FLOWERING CHERRY	NaN	NaN	KWANZAN	N	NaN	NaN
freq	NaN	57	59	464	441	8	NaN	2496	1277	4574	...	717	4582	NaN	363	NaN	NaN	363	4662	NaN	NaN
first	NaN	NaN	NaN	NaN	NaN	1989-11-15 00:00:00	NaN	NaN	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
last	NaN	NaN	NaN	NaN	NaN	2019-04-16 00:00:00	NaN	NaN	NaN	NaN	...	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
mean	10046.986200	NaN	NaN	NaN	NaN	NaN	12.132900	NaN	NaN	NaN	...	NaN	NaN	128579.562800	NaN	2.699800	2930.277600	NaN	NaN	49.247739	-123.105449
std	5808.281532	NaN	NaN	NaN	NaN	NaN	9.310923	NaN	NaN	NaN	...	NaN	NaN	74616.981285	NaN	1.550923	2032.194667	NaN	NaN	0.020973	0.049506
min	10.000000	NaN	NaN	NaN	NaN	NaN	0.250000	NaN	NaN	NaN	...	NaN	NaN	87.000000	NaN	0.000000	0.000000	NaN	NaN	49.201366	-123.223440
25%	4893.750000	NaN	NaN	NaN	NaN	NaN	4.250000	NaN	NaN	NaN	...	NaN	NaN	63023.000000	NaN	2.000000	1400.000000	NaN	NaN	49.230902	-123.144000
50%	10080.000000	NaN	NaN	NaN	NaN	NaN	10.000000	NaN	NaN	NaN	...	NaN	NaN	130094.500000	NaN	2.000000	2600.000000	NaN	NaN	49.248583	-123.102044
75%	15060.750000	NaN	NaN	NaN	NaN	NaN	17.000000	NaN	NaN	NaN	...	NaN	NaN	191550.500000	NaN	4.000000	4000.000000	NaN	NaN	49.263816	-123.062371
max	19998.000000	NaN	NaN	NaN	NaN	NaN	182.000000	NaN	NaN	NaN	...	NaN	NaN	271025.000000	NaN	9.000000	9200.000000	NaN	NaN	49.293881	-123.022469

13 rows × 21 columns

df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 5000 entries, 0 to 4999
Data columns (total 21 columns):
 #   Column              Non-Null Count  Dtype         
---  ------              --------------  -----         
 0   Unnamed: 0          5000 non-null   int64         
 1   std_street          5000 non-null   object        
 2   on_street           5000 non-null   object        
 3   species_name        5000 non-null   object        
 4   neighbourhood_name  5000 non-null   object        
 5   date_planted        2338 non-null   datetime64[ns]
 6   diameter            5000 non-null   float64       
 7   street_side_name    5000 non-null   object        
 8   genus_name          5000 non-null   object        
 9   assigned            5000 non-null   object        
 10  civic_number        5000 non-null   int64         
 11  plant_area          4963 non-null   object        
 12  curb                5000 non-null   object        
 13  tree_id             5000 non-null   int64         
 14  common_name         5000 non-null   object        
 15  height_range_id     5000 non-null   int64         
 16  on_street_block     5000 non-null   int64         
 17  cultivar_name       2700 non-null   object        
 18  root_barrier        5000 non-null   object        
 19  latitude            5000 non-null   float64       
 20  longitude           5000 non-null   float64       
dtypes: datetime64[ns](1), float64(3), int64(5), object(12)
memory usage: 820.4+ KB

There are ~600 street addresses, ~160 species_name, 22 neighbourhood in Vancouver, and 62 genus names in the dataset. There are ~5000 data points, but missing quite a few values in date_planted. Mean diameter of a tree is ~12.2, and height is 2.7. Datatypes range from datetime64, 3 float64 type, 3 int64 type, and 12 objects.

Following columns will be dropped from the analysis: std_street, on_street, civic_number, tree_id, cultivar_name. These are dropped because they do not provide any valuable information; information is irrelevant for any sort of general data exploration.

Exploratory graphs will be made for categorical and numerical datatypes. I am hoping to explore tree diversity in Vancouver through species, plant_area, spread of species across neighbourhood, age of the tree (date_planted). I also want to explore trees' height and diameters over the years and how are trees spread across the city (ie., number of trees planted on even vs odd side of the street and neighbourhoods).

For the purpose of this data exploration, no treatment will be done to the null values.

# Dropping columns that are too unique and not relevant
df = df.drop(columns=['Unnamed: 0', 'std_street','on_street', 'on_street_block', 'civic_number', 'tree_id' , 'cultivar_name'])

Evaluate distributions of all numerical columns using repeat()

#Evalute relationship between height and diameter of the tree with its species type

numerical_columns = df.select_dtypes('number').columns.tolist()
(alt.Chart(df)
 .mark_bar().encode(
     alt.X(alt.repeat(), type='quantitative', bin=alt.Bin(maxbins=25)),
     y='count()')
 .properties(width=250, height=150)
 .repeat(numerical_columns))

• Most number of trees are between diameter of 0-10.
• Most number of trees have height range of 2.0 - 2.5
• Most number of trees are found in latitude ~49.250
• Most number of trees are found in longitude ~123.07

Next, we will explore the same quantitative data using heat maps to gain better insights in relationships.

# Scroll down on the plot to see the last row
(alt.Chart(df)
 .mark_rect().encode(
     alt.X(alt.repeat('column'), type='quantitative', bin=alt.Bin(maxbins=30)),
     alt.Y(alt.repeat('row'), type='quantitative', bin=alt.Bin(maxbins=30)),
     alt.Color('count()', title=None))
 .properties(width=110, height=110)
 .repeat(column=numerical_columns, row=numerical_columns)).resolve_scale(color='independent')

Diagonal rows in the matrix are linear as they are being plotted against itself.

We see the greatest number of tree counts have diameter <10 and height between 1 - 1.5. Plotting latitude and longitude with height and diameter on a map where we can associate location with neighbourhood will be more useful. Before we progress this, let's quickly see distribution of categorical columns

categorical_columns = df.select_dtypes('object').columns.tolist()
(alt.Chart(df)
 .mark_bar().encode(
     alt.X('count()'),
     alt.Y(alt.repeat(), type='nominal', sort='x'))
 .properties(width=200, height=150)
 .repeat(categorical_columns))

Quick insights from categorical data exploration above:

• There are hundred of species. Since top 10 - 20 species make up most of biodiversity, it would be more feasible to explore the most popular species.
• Following neighbourhoods have most number of trees planted (350+): Kensington-Cedar Cottage, Renfrew-Collingwood and Hastings-Sunrise.
• Most trees are split between being planted on ODD or EVEN side of the street. This means generally tree plantation is generally balanced on both the side of a street.
• Most of the trees are along the curb (~90%).
• Genus name, assigned, common_name and plant_area don't provide too much insightful information. We can consider removing these columns in future analysis.

Next, we will explore some quantitative data using a map visualization. This is to gain insights about which neighbourhood has higher diameter and height averages.

# Import Data
url_geojson = 'https://raw.githubusercontent.com/UBC-MDS/exploratory-data-viz/main/data/local-area-boundary.geojson'
data_geojson_remote = alt.Data(url=url_geojson, format=alt.DataFormat(property='features',type='json'))
vancouver_map = alt.Chart(data_geojson_remote).mark_geoshape(
    color = 'gray', opacity= 0.5, stroke='white').encode(
).project(type='identity', reflectY=True)

# Filter the data
median_df = df.groupby('neighbourhood_name'
                      ).median().reset_index(
).rename(columns={'neighbourhood_name':'name'})[['name',
                                                 'diameter', 
                                                 'latitude', 
                                                 'longitude']]

# Develop a chart
alt.Chart(data_geojson_remote).mark_geoshape().transform_lookup(
    lookup='properties.name',
    from_=alt.LookupData(median_df, 'name', ['diameter', 'name'])).encode(
    color='diameter:Q',
    tooltip='name:N').project(type='identity', reflectY=True)

We can see on average north west end of the city has more trees that have greater diameter than south eastern. West point grey neighbourhood that the most number of trees with greater diameter on average. Next we will develop map to visualize tree's height distribution.

# Regrouping dataset
median_df = df.groupby('neighbourhood_name'
                      ).median().reset_index(
).rename(columns={'neighbourhood_name':'name'})[['name',
                                                 'height_range_id', 
                                                 'latitude', 
                                                 'longitude']]

# Developing a map
alt.Chart(data_geojson_remote).mark_geoshape().transform_lookup(
    lookup='properties.name',
    from_=alt.LookupData(median_df, 'name', ['height_range_id', 'name'])).encode(
    color='height_range_id:Q',
    tooltip='name:N').project(type='identity', reflectY=True)

Following neighbourhoods have higher number of trees with greater tree heights on average: West Point Grey, Kitsilano, Fairview, West End, Mount Pleasant, Shaunghnessy, and Dunbar-Southlands. Generally, west end of the city have greater number of trees with height averaging ~3.0. Elsewhere, height average is ~2.0, which is quite a bit shorter.

Next, we will explore distribution of species across various neghbourhood to determine which species are more popularly found.

#map
#species
(alt.Chart(df)
 .mark_bar().encode(
     alt.X('count()'),
     alt.Y('neighbourhood_name', type='nominal', sort='x'))
 .properties(width=200, height=150).facet('species_name'))

Following specifies are found across all neighbourhood: Acerifolia X, Americana, Betulus, Campestre, Cerasifera, Euchlora X, Freemani X, Kobus, Platanoides, Rubrum, Serrulata. Using this information, i would be valuable to further explore biodiversity across each neighbourhood and average age of the tree.

Next, let's see if there is any relationship between age of the tree (year it is planted) vs its height.

#Plot number of trees planted in each year with various height range

(alt.Chart(df)
 .mark_bar().encode(
     alt.X('date_planted'),
     alt.Y('count()'),
     alt.Color('height_range_id'))
 .properties(width=800, height=550))

Bar graph above doesn't provide too much insights. Generally though, height of a tree is greater in older trees. Most number of trees planted in 2007.

Next, we will explore number of trees planted per neighbourhood.

(alt.Chart(df)
 .mark_bar().encode(
     alt.X('count()'),
     alt.Y('date_planted'),
     alt.Color('neighbourhood_name'))
 .properties(width=200, height=150).facet('neighbourhood_name'))

From the data we see following neighbourhood is relatively newer because most number of trees were planted in recent years: Sunset

Following are generally older neighbourhood as most trees were planted prior to 2000s: Hastings-Sunrise, Killarney

Based on the data, we can see generally most neighbourhood are consistent with some degree of tree plantation across the years. However, some areas like Fairview, Strathcona, and West End have been quite inconsistent. Otherwise generally on average 1-2 trees are planted every year in each neighbourhoods.

Next, we will explore height and diameter of trees through density plots.

# Density plot for height and diameter 
chart1 = (alt.Chart(df).transform_density(
     'height_range_id',
     as_=['height_range_id', 'density'])
 .mark_area().encode(
     x='height_range_id',
     y='density:Q'))

chart2 = (alt.Chart(df).transform_density(
     'diameter',
     as_=['diameter', 'density'])
 .mark_area().encode(
     x='diameter',
     y='density:Q'))

chart1 | chart2

Across Vancouver, we have trees with a variety of height range from 1 - 9 m. Most trees are concentrated with height range of 2. Whereas, diameter of the trees peak at ~12 cm for the majority of the trees. It would be interesting to evaluate popular height to diamater ratio and how that ranges over the years.

Concluding Remarks

In this document various plots were made for exploring the data, which includes distribution plots, heat map matrix for quantitative and qualitative data. Quick distirbution plots were made to explore species distribution in various neighbourhoods. To better understand height and diameter distriutions, density plots were made. In addition, bar plot showing number of trees planted each year with their height distribution was developed.

Couple things that I will consider for analysis is removing all the rows that have any null values to ensure consistent data is used across. Also, through these exploratory data analysis I am interested in diving more into relationship between height and diameter of the tree with when the tree was planted. Also, I was to gain more insights about biodiversity of tree species across Vancouver neighbourhood. I will be using both the Vancouver city maps that shows distirbution of height and diameter. For other plots, I will be modifying existing plots to provide more clear picture to audience and enables to tell a clear compelling story. For example, I will plot average diameter and height versus year they were planted to see whether older trees have higher diameter and height versus younger trees. I will also develop heat map of popular species and the neighbourhood they reside in.

Following columns will not be used in further analysis: Genus name, assigned, common_name and plant_area as they don't provide relevant information for the questions I am exploring.